Meta partners with Arm to optimise AI infrastructure and softwares

Arm and Meta have announced a multi-year partnership aimed at improving the efficiency of artificial intelligence (AI) systems across hardware and software. The collaboration will focus on optimising AI performance from data centres to on-device computing, supporting the tech company's other platforms.

The partnership builds on ongoing co-design work between the two companies, combining Arm’s energy-efficient computing expertise with the tech company's advancements in AI infrastructure and open-source technologies.

Speaking of the partnership, Santosh Janardhan, Head of Infrastructure at Meta, said, “From the experiences on our platforms to the devices we build, AI is transforming how people connect and create. Partnering with Arm enables us to efficiently scale that innovation to the more than 3 billion people who use Meta’s apps and technologies.”

Rene Haas, CEO of Arm, added, “AI’s next era will be defined by delivering efficiency at scale. Partnering with Meta, we’re uniting Arm’s performance-per-watt leadership with Meta’s AI innovation to bring smarter, more efficient intelligence everywhere, from milliwatts to megawatts.” 

As part of the collaboration, the tech company will use Arm’s Neoverse-based data centre platforms to enhance the performance and reduce the power consumption of its AI-driven systems that support content discovery and personalisation across apps like Facebook and Instagram.

The companies have also optimised the Meta’s AI infrastructure software stack for Arm architectures, including open-source tools such as Facebook GEneral Matrix Multiplication (FBGEMM) and PyTorch. These enhancements, contributed back to the open-source community, are designed to improve inference efficiency and throughput.

Beyond data centres, the partnership extends to AI software across the PyTorch machine learning framework, the ExecuTorch edge runtime engine, and the vLLM inference engine. By integrating Arm’s KleidiAI technology, these efforts aim to make AI deployment more efficient across billions of devices.

Both companies plan to continue contributing to open-source AI projects to support developers building efficient AI systems worldwide.